Source URL: https://simonwillison.net/2025/May/22/tools-in-a-loop/#atom-everything
Source: Simon Willison’s Weblog
Title: Agents are models using tools in a loop
Feedly Summary: I was going slightly spare at the fact that every talk at this Anthropic developer conference has used the word “agents" dozens of times, but nobody ever stopped to provide a useful definition.
I’m now in the "Prompting for Agents" workshop and Anthropic’s Hannah Moran finally broke the trend by saying that at Anthropic:
Agents are models using tools in a loop
I can live with that! I’m glad someone finally said it out loud.
Tags: anthropic, generative-ai, ai-agents, ai, llms
AI Summary and Description: Yes
Summary: The text reflects a discussion from an Anthropic developer conference, focusing on the concept of “agents” in AI. Hannah Moran provided clarity by defining agents as models that utilize tools in a loop, highlighting the growing importance of understanding AI agents in the context of generative AI.
Detailed Description: This text captures the essence of a conference dialogue centered on defining the term “agents,” which is frequently used yet often undefined in AI discussions. The significance of this exchange lies in clarifying a fundamental concept in AI, especially for professionals working with generative AI.
– **Definition of Agents**: The term “agents” was defined by Anthropic’s Hannah Moran as “models using tools in a loop.” This definition is essential for understanding how AI systems can interactively leverage various tools to perform tasks or achieve objectives.
– **Context**: The discussion took place during an Anthropic developer conference. This context suggests that understanding agents is increasingly vital in the development of AI applications, particularly generative AI.
– **Relevance**: The prevalence of the term “agents” in conference talks without clear definitions points to a broader need for clarity and common understanding in AI terminology. Professionals in AI security, cloud computing, and infrastructure may need to address how these agents can impact security measures and best practices.
– **Implications for AI Security**: As agents evolve, understanding their roles and definitions could influence security frameworks, necessitating new strategies for managing AI interactions with tools and resources in a secure manner.
This dialogue highlights an essential step in demystifying AI functions, which has implications for the overall governance and security protocols surrounding AI technologies.